Climate Change & the use of Modelling
Atmospheric CO2 concentration is expected to rise from its current level of
354ppm to 530ppm by the year 2050 and to 700ppm by the year 2100 (Watson et al.,
1990). Concurrently this and other changes in the concentration of the infra-red absorbing
gases in the atmosphere are expected to produce a greenhouse warming of the global surface
of 3-4oC by 2100 (Bretherton et al., 1990). Predicting how vegetation
will respond to these changes is critical to understanding the impacts of atmospheric
change on both natural ecosystems and crop growth. Experimental manipulation of the
climate and atmosphere of enclosures around plants provides some direct evidence of the
effects of these changes. However, given the global scale of the changes in atmosphere and
climate, and the diversity of natural and crop ecosystems, experiments can only provide a
tiny fragment of the information needed. Further, they can only be conducted at the scale
of square meters whereas understanding at the scale of square kilometres is needed.
Predictive models provide the only alternative, and a number of empirical mathematical
models have now been developed and deployed (review: Long & Drake, 1991;Long,
1985). The empirical approach however has obvious limitations. Whilst they may closely
mimic the behaviour inside the limits of experimental information, there can be no
certainty that their behaviour outside of these limits will mimic that of actual
biological systems. Yet, understanding climate change effects requires prediction beyond
our present information about biological systems. A solution would be the use of
mechanistic models, i.e. a model based upon key biochemical and biophysical processes.
To mathematically describe the biochemistry and biophysics of all major physiological
processes in higher plants is obviously unrealistic at the present time. Variation in some
of these physiological processes is uncertain whilst for others the biochemical mechanisms
remain uncertain in any plant. However, in considering both the response of plants to
rising atmospheric CO2 concentration and the carbon balance of plants, the
process of photosynthesis is central. Photosynthesis, is the process by which plants both
sense and respond to change in atmospheric CO2 concentration. It is the
physiological process that governs the input of C to any model of plant, crop or ecosystem
production and carbon balance, and thus forms the front-end to any model concerned with
production or carbon flow. Finally, it is well suited to mechanistic modelling. With just
two variations on the basic metabolic pathway, C3 photosynthesis, the
biochemical pathway and enzymes of photosynthetic carbon metabolism are identical across
all plants. Further, Farquhar & von Caemmerer (1980) have shown from theory, that at
steady-state the responses of leaf photosynthetic CO2 uptake to light,
temperature and ambient CO2 concentration may be described by the biochemical
properties of just two steps in the process, the carboxylation reaction and the
regeneration of the acceptor for carboxylation. The kinetic properties of the primary
carboxylase, Ribulose 1:5 bisphosphatase carboxylase/oxygenase (Rubisco) are also known to
have been highly conserved in the evolution of terrestrial plants so that the broad
properties of this model should be applicable to all C3 vegetation. This
mechanistic model, developed from theory, has been widely validated as an accurate
predictor of photosynthetic carbon uptake by leaves with variation in environmental
conditions (reviewed Long, 1985). Thus, whilst all processes underlying C-balance of
vegetation cannot be mathematically described at the biochemical level, photosynthesis the
process central to the response of C-balance to atmospheric change can be mechanistically
modelled. This mechanistic model of leaf photosynthesis may then be combined with physical
models of light and gas transport within canopies to scale from the leaf to vegetation
(Long & Drake, 1991; Forseth & Norman, 1991). A few models designed to examine
atmospheric change response have already incorporated these mechanistic principles, i.e.
MAESTRO (Wang & Jarvis, 1990) and BIOMASS (McMurtrie et al, 1992). However
these models were developed for forest stands and were developed for the specialist
modellers rather than for non specialist user groups.
